On nonintrusive monitoring of electrical appliance load via restricted Boltzmann machine with temporal reservoir

Megumi Fujita, Yu Fujimoto, Yasuhiro Hayashi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

This study proposes a nonintrusive appliance load monitoring framework for estimation of the power consumptions of individual residential appliances by using aggregated total consumption based on Gaussian-softmax restricted Boltzmann machine with temporal reservoir. The proposed method is expected to estimate the hidden states of the appliances well by coding the current situation in the nonlinear temporal dynamics of the power consumption in the reservoir units so as to estimate the appliance-wise consumptions well. The accuracy of the proposed framework is evaluated by using real-world power consumption data sets.

Original languageEnglish
Title of host publicationICAART 2020 - Proceedings of the 12th International Conference on Agents and Artificial Intelligence
EditorsAna Rocha, Luc Steels, Jaap van den Herik
PublisherSciTePress
Pages902-909
Number of pages8
ISBN (Electronic)9789897583957
Publication statusPublished - 2020
Event12th International Conference on Agents and Artificial Intelligence, ICAART 2020 - Valletta, Malta
Duration: 2020 Feb 222020 Feb 24

Publication series

NameICAART 2020 - Proceedings of the 12th International Conference on Agents and Artificial Intelligence
Volume2

Conference

Conference12th International Conference on Agents and Artificial Intelligence, ICAART 2020
Country/TerritoryMalta
CityValletta
Period20/2/2220/2/24

Keywords

  • Gaussian Mixture Model
  • Gaussian-Softmax Restricted Boltzmann Machine
  • Nonintrusive Appliance Load Monitoring
  • Reservoir Computing

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software

Fingerprint

Dive into the research topics of 'On nonintrusive monitoring of electrical appliance load via restricted Boltzmann machine with temporal reservoir'. Together they form a unique fingerprint.

Cite this